What’s up with all those equals signs anyway?
66 points by ehamberg
66 points by ehamberg
This has been a little frustrating to me. I like coffeezilla and have been following his coverage of this topic on his secondary channel called voidzilla.
As soon as I've done this, I fell down the youtube rabbit-hole of mild conspiracy theorists. These other youtubers are often suggesting that the equals signs are some type of code which "the elite" uses to communicate between each other, and I'm here yelling at my screen "Dude! It's just mojibake!"
it does occur to me that a few years ago if one had said:
There are lavish parties with trafficked underage prostitutes taking place on a private caribbean island. It's run by a billionaire racist jew who calls everyone else 'goyim'. He's friends with Noam Chomsky, and the he's often visited by an Israeli Prime Minister. US Presidents Clinton and Trump have both flown in his private jet
You not only would have been called a conspiracy theorist - in many places you would have had the police at your door.
This is the kind of article I enjoy. Upon seeing the equals signs, I knew it somehow had to do with quoted-printable encoding, but couldn't for the life of me figure out how it ended up with missing characters instead of the more usual failure mode of a raw barf of encoded characters.
Maybe there needs to be a rule of the internet along the lines of "any encoding that lets you put unstructured text somewhere will evolve to contain structured text".
There are a few layers to MIME and this is one of the lower ones.
Content-type is the top level, which is what allows you to put structured text in mail messages.
The middle layer is multipart messages which support attachments and alternative encodings (eg parallel text/plain and text/html parts). HTTP and HTML can use this feature for forms (when they aren’t URL-encoded).
The lower layer is content-transfer-encoding which gets around the lack of 8-bit-clean transport for email. This can be base64 or quoted-printable, and extensions allow 8-bit and binary but I’ve never seen them used in the wild. HTTP instead splits content-transfer-encoding into two layers, content-encoding (mainly used for compression) and transfer-encoding (for HTTP/1.1 chunking), and doesn’t have to worry about 7-bit restrictions or line length limits.
This article is about quoted-printable, which is an encoding for unstructured text that contains occasional non-ascii characters and perhaps over-long lines. Quoted-printable is supposed to allow users to roughly read an encoded message even if they have a pre-MIME mail reader that doesn’t decode it. It’s designed to work with ISO 8859 western european latin alphabets, and works OK with western european text encoded in UTF-8 because of UTF-8’s backwards compatibility. So quoted-printable is over-fitted to the mostly american/european internet of the early-mid 1990s.